AI Solutions Engineer

Managed Health Care Associates, Inc. provides care communities with healthcare services, technology solutions, and insights to operate more effectively. The AI Solutions Engineer will lead the company’s agentic AI transformation by building AI engineering tooling, secure CI/CD pipelines, evaluation frameworks, governance controls, and integrations across Azure, GitHub, M365, and enterprise platforms. The role also establishes standards, enables engineering teams, and advances responsible AI adoption while safeguarding regulated data.

Responsibilities:

• Establish and continuously refine the “agentic harness,” the integrated toolchain of GitHub, VS Code, GitHub Copilot, and Claude, that full-stack Digital Solutions Engineers use for AI-assisted, agentic development
• Design and document agent orchestration patterns, including multi-agent and subagent delegation, MCP (Model Context Protocol) tool integration, and context engineering practices, that let engineers reliably delegate implementation, testing, and review work to AI agents
• Partner with squads to increase agentic code check-in rates in support of MHA's year-end adoption target, tracking and reporting progress through the METL framework
• Serve as the internal subject-matter expert on agentic coding practices, staying current on emerging AI development tools, MCP servers, and agent frameworks, and rapidly evaluating them for fit within MHA's Azure-native environment
• Design agentic workflows and tool integrations with interoperability in mind, ensuring agents, MCP servers, and enterprise systems (Jira, Azure DevOps, and platform APIs) orchestrate cleanly across squads rather than functioning as disconnected point solutions
• Build and own MHA's evaluation framework for agentic and AI-generated output, including automated evals, regression test suites, and human-in-the-loop (HITL) checkpoints for higher-risk agentic changes, closing what is currently one of the org's largest technical capability gaps
• Define quality bars and pass/fail criteria for agentic contributions prior to merge, working alongside the SonarQube and Playwright gates already embedded in the pipeline
• Ensure agentic workflows and AI tool usage comply with MHA's responsible AI use policy and applicable data privacy requirements, including safeguarding PHI and other regulated data from unintended exposure to AI tools or models
• Partner with IT Operations & Security on data-handling guardrails (masking, redaction, access scoping) for any agentic workflow that could touch sensitive member, customer, or PHI-adjacent data
• Define and enforce AI-specific security controls, including prompt-injection defenses, agent permission scoping, sandboxing of agent execution environments, and tool/action allowlisting
• Assess and mitigate secrets-exposure and software supply chain risk introduced by AI-generated code, third-party MCP servers, and agent tool integrations
• Partner with Security to fold these controls into the CI/CD pipeline and into the monthly Architecture Review Board (ARB) review process
• Build out and harden MHA's CI/CD pipelines across GitHub Actions and Azure DevOps/Azure, embedding automated security scanning (SAST/DAST, dependency and secret scanning) and quality gates directly into the pipeline
• Integrate SonarQube, Playwright, and related tooling so that code quality, test coverage, and security posture are enforced automatically, shifting quality left rather than relying on a separate QA handoff
• Define and monitor pipeline health metrics (build reliability, deployment frequency, defect escape rate, mean time to remediate) that feed the Operational Excellence and Delivery & Enablement domains
• Stand up telemetry and tracing for agentic workflows, including token spend, latency, and success/failure rates, to give engineering leadership visibility into agent and model performance
• Implement token spend and cost monitoring alongside model routing/fallback strategies that balance output quality against cost
• Feed observability data into scorecards and ARB reviews to support data-driven decisions on tool and model selection
• Work directly with Digital Solutions Engineering, Data, Analytics & Engineering, IT Operations & Security, and Agile Delivery Management to embed agentic practices, standards, and process improvements across their teams, serving as SME and standard-setter rather than as the delivery owner for their squads' work
• Develop lightweight internal training, playbooks, and office hours, favoring low-cost channels such as Microsoft Learn, to build engineering-wide fluency in agentic tools
• Identify and prototype novel, high-leverage applications of AI across Claude, OpenAI, and Azure AI services to accelerate product development and business-facing solutions, prioritizing by measurable impact rather than novelty
• Act as a thought partner to the Director of Solution Engineering and the SVP Technology on the technical roadmap for the Agentic SDLC program and represent AI engineering considerations in team level planning and Program Increment (PI) execution

Requirements:

• 5+ years of professional software engineering experience, including hands-on delivery in cloud-native environments
• Demonstrated depth in agentic AI development, including prompt engineering, agent orchestration, subagent design, MCP (Model Context Protocol) tool integration, and context engineering, not just familiarity with AI code-completion tools
• Experience building or applying evaluation frameworks for AI/agent output, such as automated evals, regression testing for agentic changes, or human-in-the-loop (HITL) review processes
• Working knowledge of AI-specific security practices, including prompt-injection mitigation, agent permission scoping, sandboxing, and tool/action allowlisting
• Strong working knowledge of Microsoft Azure and the M365 ecosystem, including Azure DevOps or GitHub Actions for CI/CD
• Proven experience building or hardening CI/CD pipelines with embedded security scanning and automated quality gates (e.g., SAST, OWASP-aligned practices, SonarQube, Playwright or equivalent test automation)
• Solid full-stack engineering foundation, comfortable across application code, infrastructure-as-code, and release automation
• Strong software engineering fundamentals independent of AI tooling: clean and maintainable code, thoughtful API and system design, robust automated testing (unit, integration, and end-to-end), and disciplined version control and code review practices
• Experience designing for interoperability and orchestration across distributed systems and APIs, ensuring agents, MCP tools, and enterprise platforms integrate cleanly rather than as isolated point solutions
• Understanding of data privacy and responsible AI obligations in a regulated, healthcare-adjacent environment (e.g., HIPAA/PHI handling), with the judgment to ensure agentic tools and workflows don't expose protected or sensitive data
• Excellent communication skills, with the ability to translate emerging AI capabilities into practical, adoptable practices for engineers who may be new to agentic workflows
• Experience with LLMOps or agent observability tooling, including tracing, token spend monitoring, and model routing/fallback, in a production setting
• Experience within a healthcare, GPO, or regulated enterprise technology environment, including direct exposure to PHI-handling or HIPAA compliance requirements
• Familiarity with Jira and modern agile delivery tooling used to plan, track, and orchestrate engineering work across squads
• Exposure to agile scaled-delivery models (SAFe-style Program Increments, team-based operating models) and performance frameworks
• Prior experience standing up an AI or platform engineering capability from the ground up, ideally as an early or founding technical hire
• Working knowledge of observability and monitoring tooling (Azure Monitor, Application Insights) and delivery/issue-tracking platforms (e.g., Jira)